conftrace_

2016 AISTATS AISTATS 2016

Distributed Multi-Task Learning

Abstract

We consider the problem of distributed multi-task learning, where each machine learns a separate, but related, task. Specifically, each machine learns a linear predictor in high-dimensional space, where all tasks share the same small support. We present a communication-efficient estimator based on the debiased lasso and show that it is comparable with the optimal centralized method.

📈 Trend Setter - Federated Learning

🐣 Hot Topic Early Bird - high-dimensional regression

Authors

Jialei Wang, Mladen Kolar, Nathan Srerbo

Topics

Artificial Intelligence > Learning Paradigms > Federated Learning Machine Learning > Core Methods > Regression

Keywords

multi-task learning feature selection distributed learning high-dimensional regression

More from AISTATS 2016

All AISTATS 2016 papers →

Scalable MCMC for Mixed Membership Stochastic Blockmodels

Non-Stationary Gaussian Process Regression with Hamiltonian Monte Carlo

A Deep Generative Deconvolutional Image Model

A Fixed-Point Operator for Inference in Variational Bayesian Latent Gaussian Models

Learning Probabilistic Submodular Diversity Models Via Noise Contrastive Estimation

Fast Saddle-Point Algorithm for Generalized Dantzig Selector and FDR Control with Ordered L1-Norm